First is the problem of alternatives. As Haswell points out, these machines fill a need. Because we have acquiesced control over the development and use of machines in writing, Haswell says we cannot blame makers of AI software, which simply “filled a vacuum we abandoned” (74). He says that the answer is that we must develop our own machines. In 1992, Paul LeBlanc expressed concern that in computers and composition, one of the richest developers of software, the practitioner/researcher, was being driven from the field. We can rant as angrily as we like, but as long as the current political structure of composition persists and as long as technological and industry development are discouraged as heavily as they are, we will have trouble resisting the use of machines. This discouragement is real; I know this from my first-hand experience with--and here I humbly and somewhat reluctantly reveal myself--Subjective Metrics, Inc., a company I co-founded to develop a software assessment tool: Waypoint. If after reading the previous sentence you are thinking I am a corporate shill, then that is exactly the problem I am talking about. At the end of their introduction, Ericsson and Haswell say they hope to start a “different revolution,” one “in which people most affected by this particular ‘emerging technology’ have a say in that future” (7). I hope we take that call seriously, but even in this book, too many others elide this problem.
|
"In 1992, Paul LeBlanc expressed concern that in computers and composition, one of the richest developers of software, the practitioner/researcher, was being driven from the field." |
Another problem is that millions of people use AI–based tools and presumably, many of them are happy, but we do not hear much from them. The result is a one-sidedness that makes me nervous about the book’s hope of expanding its audience. Ericsson and Haswell’s goal is “to better shape… the way this new instructional technology will be used at all levels” (2). The book is clear: These machines do not do what they claim. But if we are the only audience, don’t we already know (or at least suspect) that? Our critiques must affect the broader audience of users, students, parents, and policy makers, so that when we talk about problems in the structure of writing instruction, we are not just talking to ourselves. Our communications must extend beyond the eyes and ears of those who already agree with our mission, whether that involve writing to local newspapers or standing up and being heard at local school board meetings or being more aggressive in our campus communities about the degradation of teaching these machines potentially represent (Broad says "... we need to emphatically argue exactly how human instructor-evaluators provide superior educational experiences over the unarguably cheaper, faster computerized evaluation" [231]).
A final problem is that people often teach and assess writing under, shall we say, unfair circumstances. If machines offer relief from the psychological, economic, and physical strains of those circumstances, can we justify not using them, despite their limitations? Anne Herrington and Charles Moran find that although elites are not using placement technologies, many community colleges are. Some people have to.